Join the discussion
Write your take first — we'll ask for email only when you're ready to publish.
- Hacker News
- Is this the largest ever open weight model release by parameter count? I think it is.by simonw
- KIMI K3 was the biggest open weight release afaik; It is 2.8T-A100B if I'm correct
- No, Kimi k3 is 2.8T params. This is 2.4T params but ~5TB weights because it was released in bf16 and ~2.5TB for the fp8 version. Kimi k3 launched with QAT 4bit, so ~1.5TB weights.
- That's a really cool hamster [0], unfortunately it's really expensive now, 2x more expensive than Grok 4.6[1].
[0]: https://aibenchy.com/compare/x-ai-grok-4-6-high/bytedance-se...
[1]: https://aibenchy.com/compare/x-ai-grok-4-6-high/bytedance-se...
by XCSme - Interestingly, the high variant does a lot worse and failed to generate a valid SVG (and the low variant use more tokens than the high one, so maybe their reasoning efforts are not working properly).
The solar system animation is also the coolest looking I've seen, unfortunately the animation doesn't work:
https://aibenchy.com/compare/qwen-qwen3-8-2-4t-a95b-low/qwen...
by XCSme - I've been wanting to run open weight models lately to give them a shot with OpenCode. However, I get the impression that models like Qwen and Kimi k3 are impossible to run locally? I have a RTX 5090 and 64 GB of RAM but the models seem to be much larger than that. What's the route to start using these models? Bedrock?by cautiouscat
- Fireworks or OpenCode Goby verdverm
- 5090 is plenty for the Q4_K_M quantized version of 3.6 27B with reduced context size.
I run it on a 3090(24GB) and 64k context using GGUF format and llama-cpp. Double 3090 gives you 128k, quad 3090 gets you to full context - 256k.
by gessha - You could easily run any of their 30B-or-less models which is what most people are waiting for.
Apparently the ~30B variant will be released on Friday?
by zeeveener - OpenRouter is (roughly) a single proxy between you + many different models + providers. it works with opencode (+ many other products), and is relatively convenient for trying out a bunch of models.
for example, they already have qwen3.8-max
https://openrouter.ai/discover?model=qwen/qwen3.8-max
note that they add some fee ontop of things (maybe 10% of spend?). it isn't htat big of a deal for general experimentation, but if you end up wanting to use a single model in a higher-volume way, it likely makes sense to cut them out of your stack.
by mswphd - Bedrock seems to have stopped adding new open-weights models, and mostly only has Anthropic and OpenAI stuff now. You can get Qwen 3.8 directly from Alibaba: https://www.qwencloud.com (proprietary variant) or from DigitalOcean (this variant, probably also from others soon).
On your 5090 you could easily run a smaller model like Qwen 3.6 27B: https://huggingface.co/collections/Qwen/qwen36 or Gemma 4 etc., or as mentioned there's a Qwen 3.8 27B coming out in a few days.
- Vocabulary size ~248k. A bit bigger than other recent Chinese models (Kimi K3 ~164k, DeepSeek-V4 ~129k, and GLM-5.2 ~155k).
Make of this what you will.
by frozenseven - Does this mean its tokenizer is somehow tuned?by miohtama
- > Make of this what you will.
I'm interested in your take on it. IIRC Gemma family models too have a ~250k vocabulary size
- I'll just fire that up on my Intel n100...by boutell
- Not seeing the upside versus K3 here, especially with the intentional capability loss.
Read the room, Qwen. It's not a good time to hobble your releases.
by CamperBob2 - Shall we bet on when the hardware needed for this (without quantizing and at good speed) will reach < 10k USD? I'm betting 2040. I can download it now, and then get the hardware later. Eventually we can all have these things running 24/7 in our home if we wanted to. I currently would not have any task for it that would really utilize the hardware 24/7, but maybe in 20 years I will.by edg5000
- approximately $20 million for 750TB unified memory custom interconnect right nowby arthurcolle
- I would say 5 years.
The whole industry is now pushing through memory.
In 5 years you have either some type of explosion which willjust make all the hardware from today affordable or you have such an AI explosion, that the today hardware is written off and not efficient enough anymore that you can buy it for cheap.
In parallel, its clear that we need more memory.
In parallel models in hardware will become a thing on mass market.
In parallel everything gets more efficient. The 30B parameter model will be for sure more intelligent in 5 years than it is today.
- I think it is more likely that a smaller model (<400B) with similar intelligence gets developed long before the hardware to serve a 2.4T model gets cheaper than 10k.by ak_t
- More curious about how qwen3.8-27B performs. That's the size that I can run locally.by ByteWarden
- Yeah, I must have misread the press release last week as I thought it would be released at the same time.by gilgoomesh
- > In particular, Qwen3.8-Max is the official version based on Qwen3.8-2.4T-A95B with more features, such as vision input & non-thinking support, 1M context length by default, official built-in tools, etc.
That is unfortunate, that the open weight model doesn't have vision support or the 1M context length...
by l72 - Qwen3.5 was awesome: fairly open and fully featured. 3.8 lacking vision, nerfing thinking modes, and low context length feels pointless.by mips_avatar
- People have had surprising success adding vision to open-weight LLMs that ship without it, like DSV4 Flash [1] or GLM-5.2 [2]. Given this model is already vision-trained I expect that approach will work well here.
[1] https://old.reddit.com/r/LocalLLaMA/comments/1vl6ior/i_gave_...
by wren6991 - https://unsloth.ai/docs/models/qwen3.8
The 1bit quant model is at an astonishing 397GB with 95B active per MOE. This literally puts Opus 4.5 performance level into a machine a normal person could buy, and still gets usable tokens/second.
The full lossless model BF16 is clocking at 4.9TB. The model card claims the model to be between Opus 4.8 and Fable 5. Again that's astonishing as getting a machine with 7TB RAM (with context + KV cache) is still within the realm of medium size companies.
Bad things: The open source version has its vision capability removed, and the context capped at 250k . I expect someone to bolt a Kimi 2.6 vision tower to it to restore the vision capability (at less performance of course). For context, I played around with extending the context to 600k for Qwen 3.5 397b, and the context remained stable up to around 480k. It'd be interesting to see if the same can be done to Q3.8 .
Also no out of the box DSpark/DFlash support. MTP is present so we should at least get some boost in TP speed.
by guardiangod - > The 1bit quant model i
at this kind of quantization is it useful though?
by ekianjo - Opus 4.5, even 4.6-level performance has been around since July 31st in 284B total params and just 160GB of weights at native FP4 quantization- DSv4 Flash.by wolttam
- Opus 4.5 level of performance is also accessible with deepseek-v4-flash-0731 (0731 being the july 31 update) which is much, much, much smaller. 2x RTX pro 6000 blackwell can run it. 4x can run it very comfortablyby auspiv
- I don't understand the logic behind model sizes and quantization.
Suppose I have 100GB of unified memory, how should I know which model suits it best? I understand how a 2.4T model wouldn't fit, but I don't understand the impact of quantization and whether I should use a 200G model quantised to fit say 90GB of memory, or a non-quantised 90G model.
by pil0u